50 Commits

Author	SHA1	Message	Date
Reuben Morais	bd8b96c19d	Remove unneeded Saver instances	2020-02-17 15:55:16 +01:00
JRMeyer	5bba9ea5d1	Transfer-learning support	2020-02-17 08:29:10 +01:00
Reuben Morais	1e2eb96248	Update all API consumers	2020-02-11 19:44:36 +01:00
Daniel	8cc91fafb2	Moved summary printing to samples printing.	2020-02-06 14:44:41 +01:00
Daniel	63a07e6834	Added summary to evaluate_tflite.py and moved method to evaluate_tools.py.	2020-02-06 11:39:31 +01:00
Daniel	a0b5d3e7e0	Restore order of imports.	2020-02-06 11:07:11 +01:00
Daniel	320e815bb7	Remove semicolon.	2020-02-06 11:05:15 +01:00
Daniel	272ed99d24	Add median examples. Fix sorting.	2020-02-06 10:55:48 +01:00
Daniel	a2f05ccabe	Print best and worst results in a WER report.	2020-02-05 17:52:10 +01:00
Alexandre Lissy	ff401732a3	Enforce CTC decoder version check ... Fix #2710	2020-02-03 10:24:21 +01:00
juandspy	a0e528f52e	Update evaluate.py ... The package ```sys``` was not imported and gives a ```NameError: name 'sys' is not defined``` in line 96 when no test_file was provided. I added the import statement.	2020-01-21 09:57:33 +01:00
Reuben Morais	44a605c8b7	Merge pull request #2435 from mozilla/uplift-utf8-fixes ... Uplift general fixes from UTF-8 work	2019-10-25 09:09:48 +00:00
Reuben Morais	68251d6944	Avoid using references to the same object in sparse_tuple_to_text	2019-10-24 10:17:18 +02:00
Reuben Morais	d35107acdb	Merge pull request #2453 from mozilla/expose_cutoff ... Expose cutoff_prob and cutoff_top_n as flags	2019-10-23 12:21:36 +00:00
Reuben Morais	f0688ec941	Add Alphabet.encode analog to .decode and better encapsulate implementation details	2019-10-23 11:22:56 +02:00
Reuben Morais	12baf5ffbc	Expose cutoff_prob and cutoff_top_n as flags	2019-10-23 11:15:23 +02:00
Reuben Morais	739841d731	Respect FLAGS.load in evaluate.py	2019-10-14 21:03:14 +02:00
Reuben Morais	4302a5f767	Make language model scoring optional in Python inference code	2019-09-30 11:43:00 +02:00
Bernardo Henz	49c6a9c973	adding 'train_phase' to create_dataset. Now we can augment only the training-set.	2019-09-09 12:07:51 +02:00
Reuben Morais	670e06365e	Run tf_upgrade_v2 on our code	2019-08-28 17:53:24 +02:00
Reuben Morais	24bcdeb3d6	Switch from deprecated tfv1.app to absl-py	2019-08-28 10:55:33 +02:00
Tilman Kamp	007e512c00	Fix #2180 - Added wav_filename to WER report	2019-07-23 16:36:10 +02:00
Reuben Morais	6afd96e30f	Use static batch size whenever it's known	2019-07-22 14:29:45 +02:00
Reuben Morais	6566299adf	Revert "Remove deprecated SessionConfig in favor of just using defaults" as it inadvertently disables soft placement of ops ... This reverts commit dc78f8d1e6.	2019-07-15 15:03:23 +02:00
Reuben Morais	6f3e824ef7	Use tf.compat.v1 to silence deprecation warnings and enable TF 2.0 testing	2019-07-08 18:56:59 +02:00
Reuben Morais	dc78f8d1e6	Remove deprecated SessionConfig in favor of just using defaults	2019-07-08 18:56:59 +02:00
Reuben Morais	699e4ebcd7	Revert to a pipelined approach for test epochs to avoid CPU OOM with large alphabets	2019-05-13 23:49:14 -03:00
Reuben Morais	904ab1e288	Centralize progress logging and progress bar logic	2019-04-16 11:06:26 -03:00
Reuben Morais	911a1ce4b1	Do separate test epochs if multiple input files are specified	2019-04-16 11:06:26 -03:00
Reuben Morais	13757a4258	Fix pylint warnings	2019-04-11 07:02:21 -03:00
Tilman Kamp	a0c0918e25	Lazy-create global_step	2019-04-09 15:08:09 +02:00
Tilman Kamp	42e5d78e9a	Fix #2020 - Testing best-dev checkpoint	2019-04-09 15:08:09 +02:00
Reuben Morais	cc351cd607	Clean up progress bars for unknown size datasets	2019-04-08 18:04:08 -03:00
Reuben Morais	6ab91f37ec	Don't calculate dataset size by hand, use tf.errors.OutOfRangeError	2019-04-08 16:18:15 -03:00
Reuben Morais	1cea2b0fe8	Rewrite input pipeline to use tf.data API	2019-04-02 18:31:32 -03:00
Alexandre Lissy	c0cd365544	Add TFLite accuracy estimation tool ... Fixes #1852	2019-02-12 13:03:20 +01:00
Reuben Morais	f3613da82a	Use tf.contrib.layers.dense_to_sparse instead of util/ctc.py	2019-02-04 09:19:48 -02:00
Reuben Morais	7a14bcc4de	Clean up and split TensorFlow deps of text.py	2019-02-04 08:35:43 -02:00
Reuben Morais	dd6938f7af	Fix handling of Alphabet around evaluate.py	2018-12-15 09:59:49 -02:00
Reuben Morais	0c6e695138	Don't overwrite loss tensor with loss value	2018-11-13 19:16:35 -02:00
Reuben Morais	0dcb1f87c5	Clean up evaluate.py	2018-11-11 18:04:00 -02:00
Reuben Morais	5cb1aff531	Rename config singleton from C to Config	2018-11-11 18:03:52 -02:00
Reuben Morais	a5bcecbe40	Address more review comments	2018-11-11 15:24:31 -02:00
Reuben Morais	c65c22fe31	Move globals handling out of util/coordinator.py	2018-11-09 21:08:03 -02:00
Reuben Morais	56dc024d29	Centralize WER report code into evaluate.py, call it from DeepSpeech.py	2018-11-09 00:26:50 -02:00
Reuben Morais	a9e1ffc0fc	Use ctcdecode bindings in evaluate.py	2018-11-03 12:07:54 -03:00
Reuben Morais	e00c63c0a8	Fix evaluate.py to handle compact format in cached features	2018-10-11 15:58:37 -03:00
Reuben Morais	44e502e236	Feature caching	2018-09-14 11:14:36 -03:00
Reuben Morais	51cac1031b	Address review comments	2018-08-03 14:46:05 -03:00
Reuben Morais	831ebf4081	Add evaluate.py script to create a WER report from a checkpoint using the inference graph	2018-08-02 13:22:24 -03:00